Red Cell Distribution Width-to-Albumin Ratio as a Predictor of Preeclampsia in Advanced Maternal Age: A Retrospective Cohort Study

Introduction

Preeclampsia is a pregnancy‑specific hypertensive disorder characterized by new‑onset hypertension and end‑organ dysfunction after 20 weeks of gestation.^1,2 It remains one of the leading contributors to maternal and perinatal morbidity and mortality worldwide, despite advances in obstetric care.² Current evidence indicates that abnormal placentation, systemic endothelial injury, oxidative stress, and an exaggerated maternal inflammatory response are key pathological features of the disease.^3–5

The growing proportion of pregnancies in women of advanced maternal age (AMA, commonly defined as ≥35 years) has important implications for obstetric risk.^2,6 AMA is associated with higher rates of chronic hypertension, diabetes, obesity, and subclinical vascular dysfunction, all of which predispose to preeclampsia.^2,6,7 Nevertheless, early identification of high‑risk AMA women in routine clinical practice remains challenging because available screening tools are either resource‑intensive or not widely implemented in low‑ and middle‑income settings.^8,9

Red cell distribution width (RDW) reflects heterogeneity in erythrocyte size and is influenced by inflammation, oxidative stress, and impaired erythropoiesis.^10–12 Elevated RDW has been linked to adverse outcomes in cardiovascular disease and critical illness.^13,14 Serum albumin, conversely, is a negative acute‑phase reactant with antioxidant and endothelial‑stabilizing properties;^15,16 hypoalbuminemia is often observed in systemic inflammation and may reflect both nutritional status and capillary leak.¹⁵ Importantly, RDW has been associated with inflammation and oxidative stress and reflects erythrocyte size heterogeneity,¹⁰ whereas serum albumin reflects nutritional status and systemic inflammation and may also relate to endothelial permeability.¹⁵ RAR, derived from RDW and serum albumin and based on routinely available laboratory parameters, may serve as a simple and readily available biomarker, although its role in preeclampsia risk assessment remains to be established.^17,18 Given that preeclampsia is characterized by systemic inflammation, endothelial dysfunction, and impaired vascular adaptation, RAR may be of potential interest in this context.^3,4,17

The red cell distribution width‑to‑albumin ratio (RAR) integrates these two routinely measured parameters into a single composite index.^17,19 Recent studies in intensive care settings suggest that a higher RAR is associated with adverse outcomes and increased mortality in patients with sepsis, acute pancreatitis, and rheumatic diseases.^18–20 Population‑based cohort data further indicate that elevated RAR is linked to higher all‑cause and cardiovascular mortality in the general population,¹⁷ highlighting its potential as a simple marker of systemic inflammation and vascular stress. RAR has been studied in non-obstetric inflammatory and cardiovascular conditions,^21,22 but evidence in obstetric populations, especially for preeclampsia, remains limited. To our knowledge, however, no prior study has examined the relationship between RAR and preeclampsia, particularly in women of advanced maternal age.

Given the central roles of inflammation, endothelial damage, and microcirculatory impairment in preeclampsia, we hypothesized that higher RAR values would be associated with an increased risk of preeclampsia among women of advanced maternal age.^3–5 In this retrospective cohort study, we aimed to: (1) examine the association between RAR (analyzed as RAR×10 and by tertiles) and preeclampsia; (2) explore potential non‑linear dose–response relationships; and (3) assess whether this association differs across clinically relevant subgroups.

Materials and Methods

Study Design and Population

We conducted a retrospective cohort study at the First People’s Hospital of Changde City, a tertiary grade A hospital in Hunan Province, China. Electronic medical records of pregnant women who delivered between January 2013 and December 2023 were reviewed. The study protocol was reviewed and approved by the Academic Ethics Committee of Changde Hospital, Xiangya School of Medicine, Central South University (First People’s Hospital of Changde) (approval No. 2024–190-01). Given the retrospective nature of the study and the use of routinely collected clinical data, the requirement for informed consent was waived by the ethics committee. All data were anonymized/de-identified before analysis. The study was conducted in accordance with the Declaration of Helsinki. Eligible participants were women who (1) were aged ≥35 years at the time of delivery; (2) had singleton pregnancies; and (3) had complete baseline laboratory tests during pregnancy, including RDW and serum albumin. We excluded women with missing data on glucose, triglycerides, weight, or height (n = 66); those with gestational diabetes mellitus (n = 20); those with severe hepatic or renal dysfunction (n = 12), malignant tumors or autoimmune diseases (n = 8); and those with pre‑existing chronic hypertension before pregnancy or diagnosed before 20 weeks’ gestation (n = 142). After applying these criteria, 2296 women were included in the final analysis. A flowchart of participant selection and exclusion is shown in Figure 1.

Figure 1 Flowchart of participant selection and enrollment process for the study.

Exposure Assessment: Red Cell Distribution Width‑to‑albumin Ratio

During routine early pregnancy prenatal visits, typically occurring at 11–12 weeks of gestation, all women provided early‑morning fasting blood samples after 8–10 hours without food. RDW (%) and serum albumin (g/L) were then quantified in the hospital laboratory using automated hematology and clinical chemistry platforms operated under strict quality‑control procedures. The RAR calculation was based on the laboratory measurement obtained at this early pregnancy visit for each participant. As our study relied on routinely collected retrospective clinical data, only one eligible RAR measurement per participant was available for analysis.

RAR was defined as:

RAR = RDW (%) / albumin (g/L)¹⁹

To simplify interpretation and better align with the distribution of RAR values, we multiplied RAR by 10 and used RAR×10 as the primary exposure variable in all regression models and spline analyses. Participants were further stratified into tertiles according to RAR: tertile 1 (lowest), tertile 2 (middle), and tertile 3 (highest). The cut‑off values corresponded approximately to three RAR groups, as derived from the cohort distribution: Group 1 (0.240 ≤ RAR < 0.377), Group 2 (0.377 ≤ RAR < 0.419), and Group 3 (0.419 ≤ RAR ≤ 1.364).

Outcome Definition: Preeclampsia

Preeclampsia was diagnosed by experienced obstetricians according to contemporary international criteria after 20 weeks of gestation, and was defined as new-onset hypertension with a systolic blood pressure of at least 140 mmHg and/or a diastolic blood pressure of at least 90 mmHg on at least two occasions after 20 weeks of gestation in previously normotensive women. This was accompanied by either proteinuria of at least 300 mg per 24 hours or a dipstick reading of at least 1+, or, in the absence of proteinuria, new-onset maternal end-organ dysfunction such as thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or new-onset cerebral or visual disturbances.²³ Women with chronic hypertension who developed superimposed preeclampsia were excluded by design.

The primary outcome for this study was the presence of preeclampsia (yes/no).

Covariates

Information on maternal demographic, obstetric, clinical, and laboratory characteristics was retrieved from the hospital electronic medical record and laboratory information systems. Maternal variables comprised age, body mass index (BMI), gestational week at delivery, gravidity, parity, systolic and diastolic blood pressure, heart rate, gestational diabetes mellitus (GDM), family history of hypertension, documented cardiovascular diseases (CVD), liver insufficiency, and baseline abnormalities in blood coagulation profiles. Laboratory measurements comprised white blood cell (WBC) count, red blood cell (RBC) count, platelet (PLT) count, neutrophil‑to‑lymphocyte ratio (NLR), blood urea nitrogen (BUN), serum creatinine (Cr), alanine aminotransferase (ALT), aspartate aminotransferase (AST), fibrinogen (FIB), D‑dimer, triglycerides (TG), serum uric acid (UA), fasting glucose (GLU), hemoglobin (Hb), high‑density lipoprotein cholesterol (HDL‑C), and low‑density lipoprotein cholesterol (LDL‑C). Candidate variables were entered into multivariable models as covariates based on established risk factors for preeclampsia reported in the literature and on statistical considerations for model specification.

Statistical Analysis

The normality of continuous variables was assessed using the Shapiro–Wilk test. Continuous variables were summarized as mean ± standard deviation (SD) for approximately normally distributed data or as median with interquartile range (IQR) for skewed data. Categorical variables were presented as counts and percentages. Baseline characteristics across tertiles of RAR were compared using one‑way analysis of variance (ANOVA) for normally distributed continuous variables or Kruskal–Wallis tests for skewed continuous variables, as appropriate, and chi‑square tests for categorical variables. The specific statistical test used for each variable is indicated in the table footnotes. To examine crude associations between each candidate variable and preeclampsia, we first performed univariable logistic regression analyses. Odds ratios (ORs) and 95% confidence intervals (CIs) from these models are reported in Supplementary Table 1. We then applied multivariable logistic regression to quantify the association between RAR×10 and preeclampsia. An unadjusted model and two progressively adjusted models were fitted:

(1) an unadjusted model including no covariates;

(2) Model I, adjusted for age, body mass index (BMI), and parity;

(3) Model II, additionally adjusted for triglycerides (TG), cardiovascular diseases (CVD), gestational diabetes mellitus (GDM), uric acid (UA), high‑density lipoprotein cholesterol (HDL‑C), low‑density lipoprotein cholesterol (LDL‑C), and fasting glucose (GLU).

RAR×10 was analyzed both as a continuous exposure and as a categorical variable according to tertiles, with the lowest tertile serving as the reference group. A linear trend across tertiles was evaluated using a trend test. Potential non‑linear associations between RAR×10 and preeclampsia were assessed using restricted cubic spline (RCS) functions. These models were adjusted for the same covariates as in Model II. P values for the overall association and for non‑linearity were reported. Predefined subgroup analyses were conducted to evaluate effect modification by BMI (<28 vs ≥28 kg/m²), GDM status (no vs yes), gravidity (<3 vs ≥3 pregnancies), heart rate (<80 vs ≥80 beats/min), gestational week at delivery (<37 vs ≥37 weeks), platelet count (<125 vs ≥125×10⁹/L), and uric acid (<420 vs ≥420 µmol/L). Interaction terms between RAR×10 and each stratification factor were added to the fully adjusted model, and P values for interaction were obtained from likelihood ratio tests comparing models with and without the corresponding interaction term.

All statistical tests were two‑sided, and P values <0.05 were considered statistically significant. Analyses were performed using R software (version 4.3.1; http://www.R-project.org, The R Foundation) and Free Statistics software (version 2.3 beta).

Results

Baseline Characteristics Across RAR Tertiles

A total of 2296 women of advanced maternal age (mean age 37.78 ± 2.67 years) were included, of whom 328 (14.29%) developed preeclampsia. Baseline characteristics according to RAR tertiles are summarized in Table 1. Age, BMI, gravidity, parity, and systolic blood pressure all differed significantly across tertiles (all p<0.001). Metabolic and biochemical indices also varied across tertiles. Compared with tertile 1, women in tertile 3 had higher fasting glucose, triglycerides, creatinine, and D‑dimer levels, as well as a greater prevalence of GDM and CVD (all p≤0.01). Several hematologic and clinical variables also differed across tertiles, including white blood cell count, neutrophil-to-lymphocyte ratio, platelet count, liver insufficiency, and abnormal blood coagulation status (Table 1). The prevalence of preeclampsia increased steadily across RAR tertiles: 9.95% in tertile 1, 14.14% in tertile 2, and 18.75% in tertile 3 (p<0.001).

Table 1 Baseline Characteristics According to Red Cell Distribution Width-to-Albumin Ratio

Univariate Analysis for Preeclampsia

Supplementary Table 1 summarizes the univariate associations with preeclampsia. Higher BMI, systolic and diastolic blood pressure, GDM, and a family history of hypertension were all strongly associated with increased odds of preeclampsia. Several laboratory markers, including NLR, creatinine, triglycerides, uric acid, fasting glucose, and heart rate, also showed positive associations (all p<0.05). These findings guided the selection of covariates in multivariable models assessing the RAR–preeclampsia relationship.

Association Between RAR and Preeclampsia

Table 2 presents the multivariable logistic regression analyses. In the unadjusted model, each 1‑unit increase in RAR×10 was associated with 15% higher odds of preeclampsia (OR = 1.15, 95% CI: 1.05–1.26; p = 0.002). After adjustment for age, BMI, and parity (Model I), this association became stronger (OR = 1.22, 95% CI: 1.10–1.34; p<0.001). Further adjustment for TG, CVD, GDM, UA, HDL‑C, LDL‑C, and fasting glucose (Model II) slightly attenuated the effect estimate, but the association remained statistically significant (OR = 1.16, 95% CI: 1.04–1.30; p = 0.006).

Table 2 Results of Multivariable Logistic Regression Conducted to Evaluate the Relationship Between the Red Cell Distribution Width-to-Albumin Ratio and Risk of Preeclampsia

When RAR×10 was categorized into tertiles, women in the highest tertile had higher odds of preeclampsia compared with those in the lowest tertile, whereas the association for the intermediate tertile was attenuated. In the fully adjusted model, the ORs (95% CI) for tertiles 2 and 3 were 1.38 (0.98–1.94) and 1.89 (1.35–2.64), respectively. The p for trend across tertiles was <0.001, indicating a graded association between higher RAR and preeclampsia risk.

Restricted Cubic Spline Analysis

The restricted cubic spline analysis revealed a significant non‑linear association between RAR×10 and the odds of preeclampsia (p for overall association <0.001; p for non‑linearity <0.001) (Figure 2). Using the cohort mean of RAR×10 (3.95) as the reference, odds ratios were below 1.0 at lower RAR×10 values and increased steeply as RAR×10 approached the reference range, where the risk peaked. Beyond the reference value, the curve slightly bent downward, with odds ratios gradually declining at higher RAR×10 levels. Confidence intervals widened toward both extremes of the distribution, reflecting fewer observations at very low and very high RAR×10 values.

Figure 2 RCS curve analysis.

Subgroup Analyses

Figure 3 presents exploratory subgroup analyses of the association between RAR×10 and preeclampsia. The positive association was broadly consistent across most clinical strata. A statistically significant interaction was observed for BMI (p for interaction <0.001). Among women with BMI <28 kg/m², RAR×10 was not significantly associated with preeclampsia (OR = 0.91, 95% CI: 0.74–1.11), whereas in those with BMI ≥28 kg/m², each 1‑unit increase in RAR×10 was associated with approximately threefold higher odds of preeclampsia (OR = 3.46, 95% CI: 2.49–4.81). No significant interactions were detected for gravidity, heart rate, gestational week at delivery, platelet count, or GDM (all p for interaction >0.05). A significant interaction was also observed for uric acid (p for interaction =0.001), with a stronger association among women with UA ≥420 µmol/L (OR = 1.77, 95% CI: 1.28–2.45).

Figure 3 Subgroup forest plot.

Discussion

In this retrospective cohort of 2296 women of advanced maternal age, we observed a statistically significant and non‑linear association between the red cell distribution width‑to‑albumin ratio and preeclampsia. Higher RAR×10 values may be associated with increased odds of preeclampsia, and women in the highest tertile had approximately double the risk compared with those in the lowest tertile, even after extensive adjustment for established risk factors. RCS analysis suggested a J‑shaped dose–response pattern. However, the confidence intervals became wider at the extremes of the distribution, and this part of the curve should therefore be interpreted cautiously. Subgroup analyses further suggested that the association may be more pronounced in women with elevated BMI and some features of cardiometabolic stress, although these findings should be interpreted as exploratory.

To place our findings in context, prior studies in non-obstetric populations have linked elevated RAR to worse clinical outcomes, including higher mortality in critically ill patients with rheumatic diseases, a greater likelihood of type I cardiorenal syndrome in patients with acute myocardial infarction, and more severe disease in patients with acute pancreatitis.^19,24,25 Other studies have similarly suggested that RAR may reflect systemic inflammation in different disease settings.^24,26 In addition, elevated RAR has been associated with worse clinical outcomes or prognosis in several conditions, including cardiorenal syndrome, chronic kidney disease progression, and mortality among hypertensive individuals.^24,26,27 Although direct evidence regarding RAR and preeclampsia is still lacking, the direction of our findings is broadly consistent with this emerging literature. At the same time, these studies were conducted in different populations and with different clinical endpoints, so our results should be viewed as extending the existing evidence base to women of advanced maternal age rather than confirming a uniform effect across settings.

In the present study, higher RAR was associated with a greater likelihood of preeclampsia, which may be biologically plausible given the central roles of inflammation, oxidative stress, endothelial dysfunction, and placental hypoperfusion in this disorder.^1,2 One possible explanation is that the RDW component reflects anisocytosis related to chronic inflammation, oxidative stress, iron deficiency, or disturbed iron metabolism, and impaired erythropoiesis.^10,14 Increased erythrocyte size heterogeneity may alter blood flow patterns and enhance interactions with the vascular wall.²⁸ In preeclampsia, altered erythrocyte rheology and aggregation may further impair placental microcirculation and oxygen delivery, thereby contributing to placental hypoxia.²⁹ At the same time, albumin has antioxidant properties, and lower albumin levels may reflect underlying pathophysiological disturbances in preeclampsia.¹⁵ From this perspective, a higher RAR may reflect erythrocyte aggregation and rheological abnormalities observed in preeclampsia,²⁹ as well as broader pathophysiological disturbances, including inflammation, oxidative stress, and endothelial dysfunction.^1,2 Although direct evidence linking RAR to preeclampsia remains limited, obstetric studies support the relevance of its components. A systematic review and meta-analysis reported higher RDW levels in women with preeclampsia than in controls,³⁰ whereas albumin has antioxidant properties, and hypoalbuminemia in preeclampsia may be related to inflammatory activity and increased urinary albumin excretion, with a possible association with oxidative stress.¹⁵ Genetic factors may also contribute to susceptibility to preeclampsia, particularly in the context of advanced maternal age,³¹ but whether such factors affect RAR or modify its association with preeclampsia remains unclear. Thus, these mechanistic interpretations remain inferential because molecular, placental, and genetic features were not directly assessed in this study.

Our subgroup findings may be relevant to clinical practice. In particular, the much stronger association between RAR×10 and preeclampsia among women with BMI ≥28 kg/m² in our study suggests that obesity-related low‑grade chronic inflammation, insulin resistance, and endothelial dysfunction may amplify the detrimental vascular and inflammatory effects reflected by this index, in line with prior evidence that maternal obesity is characterized by systemic inflammation and vascular dysfunction and substantially increases the risk of hypertensive disorders of pregnancy, including preeclampsia.^32,33 Women with elevated prepregnancy BMI are at increased risk of preeclampsia and other adverse pregnancy outcomes, highlighting the need for increased preventive attention to weight management before pregnancy.³⁴ Obesity, characterized by low‑grade chronic inflammation, insulin resistance, and endothelial dysfunction,^32,35 may synergize with the hematologic and biochemical disturbances captured by RAR. Similarly, the stronger associations observed in women with thrombocytopenia or hyperuricemia are consistent with more advanced endothelial and microvascular injury, as thrombocytopenia is a hallmark of pregnancy‑related thrombotic microangiopathies characterized by endothelial and small‑vessel damage, and elevated uric acid levels in preeclampsia are associated with oxidative stress and systemic endothelial dysfunction in preeclampsia and related hypertensive disorders of pregnancy.^36,37 Nevertheless, these subgroup findings should be interpreted cautiously, especially because not all interaction tests were statistically significant.

Another important consideration is the potential public health implications of implementing RAR as a candidate marker for early risk assessment in clinical settings. By identifying women at higher risk for preeclampsia through early screening methods, healthcare providers could implement targeted interventions involving closer monitoring and lifestyle modifications.³⁸ This proactive approach, by facilitating early detection and targeted interventions, could significantly reduce the morbidity and mortality associated with preeclampsia, particularly within vulnerable populations.³⁹ Furthermore, lifestyle-focused educational interventions may be beneficial for pregnant women with hypertension, particularly in promoting healthy behaviors and supporting blood pressure management.⁴⁰ However, formal validation studies are still required before RAR can be recommended for routine screening use.

RAR has several advantages as an index derived from routinely available laboratory parameters. RDW and albumin are widely available, low‑cost assays that are routinely performed in hospital laboratories.^17,41 The ratio is simple to calculate and does not require specialized equipment or additional blood sampling. Compared with more complex multi‑marker panels or imaging‑based approaches, RAR could be readily implemented in resource‑constrained settings as part of early antenatal assessment.^41,42 Women of advanced maternal age are at increased risk of adverse pregnancy outcomes and may warrant closer surveillance during pregnancy.⁴³ In this context, the association observed in our study suggests that RAR may reflect underlying pathophysiological disturbances related to preeclampsia in this population. However, these findings should be interpreted cautiously, and whether RAR provides additional information beyond established risk factors requires further validation.

Our study also has limitations. First, its retrospective single‑center design precludes causal inference and may limit generalizability to other populations or healthcare systems. Second, RAR was assessed only once, using the earliest routine antenatal laboratory measurement obtained in early pregnancy, typically at 11–12 weeks of gestation; we were therefore unable to evaluate longitudinal changes or the optimal gestational window for measurement. Third, although we adjusted for multiple confounders, residual confounding by unmeasured factors (eg, dietary patterns, socioeconomic status, or inflammatory markers not captured in routine tests) cannot be excluded. Finally, our study population consisted solely of women with advanced maternal age, and the applicability of these findings to younger pregnant women needs to be confirmed.

Future research should validate these findings in multicenter, prospective cohorts with serial RAR measurements across gestation and clearly defined sampling windows. Furthermore, comparative modeling studies integrating RAR with existing clinical data would help clarify whether RAR provides incremental predictive value beyond established risk factors and biomarkers for preeclampsia risk assessment. Formal evaluation of predictive performance, including sensitivity, specificity, and calibration, will be needed before clinical implementation can be considered. Mechanistic investigations integrating RAR with direct measures of endothelial function, oxidative stress, and placental pathology would further elucidate the pathophysiological pathways underpinning the observed association.

Conclusion

In conclusion, in this single‑center retrospective cohort of pregnant women of advanced maternal age, a higher red cell distribution width‑to‑albumin ratio measured in early pregnancy was associated with a higher risk of preeclampsia. The association was non‑linear and appeared stronger in women with obesity. However, these findings do not establish the predictive performance or clinical utility of RAR. These findings warrant confirmation in prospective, multicenter studies with clearly defined sampling windows and formal prediction analyses, as well as in more diverse obstetric populations.